4D Total-Body PET Parametric Imaging with Cardiac Modulation

54 Introduction: Total-body dynamic PET can provide tracer kinetic assays of physiologically and biologically relevant information across the entire human body. Our previous work has demonstrated the capability of performing dynamic PET imaging with 100-ms temporal resolution on the uEXPLORER scanner. Sub-second dynamic PET imaging allows clear visualization of fast tracer dynamics after bolus injection with physiological (cardiac and respiratory) motion in real-time. In this work, we aim to investigate a 4D kinetic analysis method with cardiac modulation for total-body parametric imaging based on compartmental modeling of the sub-second early dynamic PET data. Methods: A total-body dynamic 18F-FDG PET study was acquired on the uEXPLORER scanner. The first 2-minutes of the scan were divided into temporal frames with two protocols, 1 s × 120 and 0.1 s × 1200, respectively. Quantitative image reconstruction was performed using our total-body kernel-based algorithm with corrections for photon attenuation, detector normalization, scatter and random events. In light of the fact that the first pass of FDG through tissues is mainly related to blood flow rather than the glucose metabolism, we adopted a one-tissue compartment method (1TC) for total-body parametric imaging. An image-derived input function was extracted using the time activity curves (TAC) from a region placed over the ascending aorta (AA). Considering the large variation of bolus arrival times across the body, a voxel-based time delayed input function was modeled and fitted relative to the AA TAC. To investigate the effect of cardiac motion on tracer kinetics, an image-derived cardiac gating signal was extracted using a band pass filtered TAC from a fixed myocardium region. Tissue TACs were phase-gated into 8 cardiac gates and the gated TAC of each voxel was fitted to the 1TC model. In addition, we used the Akaike information criterion (AIC) to select the optimal model between the 1TC models with and without perfusion parameters. Total-body parametric images were generated for eight cardiac gates. To quantify the cardiac modulation effect, regions of interest (ROIs) were drawn in the gray matter (GM) and white matter (WM). The mean and standard deviation of the estimated Vb values over 10 ROIs within GM and WM were calculated and compared. Results: Comparing the dynamic TACs of different organs and tissues at 1-s temporal resolution and 0.1-s temporal resolution, we see that the 0.1-s TACs can capture cardiac motion and variation of the blood flow within each cardiac cycle, but those effects are completely smoothed out in the 1-s TACs. Applying the band pass filter to the myocardial TAC resulted in a clean cardiac motion signal, which can be used to extract phase-gated dynamic data. The AIC difference between the 1TC without and with perfusion parameters shows that voxels containing large vessel (e.g. veins and arteries) prefer a simplified blood-only model. Total-body parametric Vb images estimated at different cardiac phases show visible difference between the end-systole and end-diastole. Furthermore, comparing Vb values in the brain regions over the 8 cardiac gates, we can see some small variations in Vb, demonstrating a potential cardiac modulation effect. This effect was observed in both GM and WM. Conclusion: We have demonstrated a 4D total-body parametric imaging method using image-derived cardiac gating on early FDG dynamic PET data. The proposed method provides total-body parametric images for studying blood volume over the cardiac cycle. The observed cardiac modulation effect may also provide additional information for studying tissue function and mechanical properties. Acknowledgements: Support for this work includes NIH grant R01 CA206187 and a UC Davis Innovative Development Award. We acknowledge the contributions of all team members from UC Davis, United Imaging Healthcare and Zhongshan Hospital.